The Data Backup Thread (& request for more suggestions)

Win7 has a nice built-in backup, including full image capability, so you can boot up from the Win7 DVD and restore a full working machine from the image. I've done it several times, and it works very well.

If you're backing up to another drive on the same machine, it'll do automatic versioning and space management; if you're backing up to a network share, it doesn't seem to do any space management, so you have to manually go in and delete old backups when you're low on space.

I saw that Win7 had that ability, but I just assumed that it was terrible and clunky. I'll give it a go.
Any experience with restoring those images to dissimilar hardware?

No, only on the same computer. You could potentially run into issues if you restored to a system drive that the OS didn't have a driver for, but otherwise I'd expect a hardware transfer to probably work okay. Just remember you'll need to reactivate it within a few days.

Putting the question in this thread because it's pretty much related.

Does anyone have any great ideas on how to verify that data is good, either in a backup or on the 'live disc'?

Assuming you start from one point with a known good set of data (as much as you can), I know there are tools like SFV to take a checksum, and you could always check against your backup (and if there was a difference you would need to determine which was good), and you need to make sure you're actually backing up good data and not damaged data. Are there any strategies to link that kind of checking with everyday usage and take into account when a file is intentionally changed?

Just wondering if anyone can give me a head start on that line of thinking.

and you need to make sure you're actually backing up good data and not damaged data.

Ok, what you're talking about there is data validation. This isn't very common yet, but will become more so as drives get bigger. Typically, this is done via a checksum. It can be implemented either at the block storage level, either with software RAID or a RAID controller, or it can be done at the filesystem level. But the only filesystem I know that presently does this and is fully production ready is ZFS, which is limited to Solaris and FreeBSD.

If you're willing to use Linux, the software mdraid is both fast and pretty sophisticated, though it will take some learning to implement.

If you want to stay on Windows, its software RAID solutions are horrible, so you'd want to move to a RAID controller. (which is very expensive, especially with the battery backup module they need to really run fast while being safe with your data.) Areca has some pretty good SAS/SATA controllers that can take either kind of drive. The controlers are typically $500-$600, the battery is another $125 or so, and then you'll need storage drives, and optionally some enclosures to put them in. You can expect to spend about $2,000 for a RAID6 setup.

I'm building a new gaming rig soon and, will be dipping my toes into the whole backup scene. I've never really paid attention to it in the past for various reasons* but, now it seems like I should at least have a snapshot of my fresh install for future use.

How much space would a drive image take up; lets say 2TB drive with Win7 Home, basic software (Microsoft Security Essentials, Chrome, Java, Flash, Dropbox, etc), drivers and, device software?

How easy are these to compress; is that wise? This will be stored in the cloud (Dropbox or Skydrive) Would/should these be secured in some fashion since they are going to non-secured cloud storage?

Will the Win7 utility be the best option for something like this?

*This will be for my gaming rig; no important info will be stored (I don't view saved games as important); reinstalling stuff and re-downloading it is a small inconvenience. When the prices on platters come back down, I will look at doing full backups then.

Tigerbill wrote:

I'm building a new gaming rig soon and, will be dipping my toes into the whole backup scene. I've never really paid attention to it in the past for various reasons* but, now it seems like I should at least have a snapshot of my fresh install for future use.

How much space would a drive image take up; lets say 2TB drive with Win7 Home, basic software (Microsoft Security Essentials, Chrome, Java, Flash, Dropbox, etc), drivers and, device software?

How easy are these to compress; is that wise? This will be stored in the cloud (Dropbox or Skydrive) Would/should these be secured in some fashion since they are going to non-secured cloud storage?

Will the Win7 utility be the best option for something like this?

*This will be for my gaming rig; no important info will be stored (I don't view saved games as important); reinstalling stuff and re-downloading it is a small inconvenience. When the prices on platters come back down, I will look at doing full backups then.

Here is my 2 cents. If you are not going to store unrecoverable data on it (ie. photos, docs, etc.) and all you want is to have a fast way to get it back to "factory" make a restore on a usb. I have been doing this and have had to do a few restores and installing an OS from a usb is super fast vs. using a cd.

My wife and I have been using Time Machine since it came out, which is all well and good if/when our hard drives explode. So we pass the *Legion* Backup Test, or what I prefer to specify as the Weak *Legion* Backup Test.

But now that we've gone all "new parent"-y, our irreplaceable photos have become no-really-they're-goddamned-irreplaceable photos. So while we're protected against exploding hard drives, what about exploding apartments? This is what I call the Strong *Legion* Backup Test: is your data safe and recoverable if your physical computer disappears/explodes/clams up this instant. That's touched on in the OP, but with the cloud becoming more and more mainstream, it's important to be explicit about it.

So that said, in looking for a brain-dead easy offsite backup solution, I've found CrashPlan+ Family—cheaper than Netflix and infinitely more valuable. CrashPlan already has the *Legion* stamp of approval, in addition to Ars' and Lifehacker's.

Any other opinions? CP looks like it does exactly what I need it to, which is complement Time Machine both in terms of being offsite, and in being equally dead simple to set up and completely automatic.

I've been using CrashPlan for a while now and am signing my girlfriend up for it to back up her multitude of photos. I love it! I used to use JungleDisk but it charges for bandwidth and storage whereas this is one price no matter how much you have. You point it to the folders and pretty much never have to open it again. I have it send me an e-mail once a week to confirm the backups are running correctly and beyond that, it just sits there and quietly works. Best and easiest cloud backup solution I've used to date.

Parallax Abstraction wrote:

it just sits there and quietly works. Best and easiest cloud backup solution I've used to date.

Thanks PA, that's just what I wanted to hear!

Parallax Abstraction wrote:

I've been using CrashPlan for a while now and am signing my girlfriend up for it to back up her multitude of photos. I love it! I used to use JungleDisk but it charges for bandwidth and storage whereas this is one price no matter how much you have. You point it to the folders and pretty much never have to open it again. I have it send me an e-mail once a week to confirm the backups are running correctly and beyond that, it just sits there and quietly works. Best and easiest cloud backup solution I've used to date.

Gravey wrote:

CrashPlan already has the *Legion* stamp of approval

Yeah, I need to edit that into the top post.

CrashPlan has become my #1 go-to recommendation for cloud backup now, both personal and at work.

It has replaced Carbonite for the backup system I make my family members install on their computers.

It (or, rather, the PRO version) also is in the process of replacing JungleDisk on our servers at work. We're in that long initial upload process, so we're continuing to run JD still. But once CrashPlan has everything backed up, we may shut the JD backup down. Or maybe continue to run in parallel (two separate cloud backups for important data isn't such a bad idea). CrashPlan PRO will definitely be the go-to for new servers, though.

Also, a brilliant feature of CrashPlan is the ability to store an encrypted backup on another CrashPlan-running machine. I have my wife's laptop backing up all her photos to an encrypted blob on her desktop machine, in addition to the cloud backup. So, if only the laptop died, we could do a fast local recovery from the backup stored on her desktop, rather than the slower recovery from the cloud backup.

And that feature is free, even if you don't pay to use CrashPlan's cloud storage.

*Legion* wrote:

Also, a brilliant feature of CrashPlan is the ability to store an encrypted backup on another CrashPlan-running machine. I have my wife's laptop backing up all her photos to an encrypted blob on her desktop machine, in addition to the cloud backup. So, if only the laptop died, we could do a fast local recovery from the backup stored on her desktop, rather than the slower recovery from the cloud backup.

What about encrypting the data going up to the cloud? I want to back up my financials off site, but not sure I want them stored in the clear in the cloud.

Mixolyde wrote:

What about encrypting the data going up to the cloud? I want to back up my financials off site, but not sure I want them stored in the clear in the cloud.

Here's an FAQ on CrashPlan's encryption.

Add me to the list of those very, very happy with CrashPlan+.

Mixolyde wrote:

What about encrypting the data going up to the cloud? I want to back up my financials off site, but not sure I want them stored in the clear in the cloud.

I won't use any cloud backup that doesn't include encryption under my control.

CrashPlan lets you define your own encryption key (or a password that gets hashed into a key) which each file gets encrypted against. That key doesn't get transmitted to CrashPlan, and they make sure you know that you will be unable to recover your stuff if you lose that key.

With encryption/security, unless you're going to do an inspection of all the code you're using, at some point you have to trust someone that it's secure.

My company is using Synctoy as a stop gap for local backups until we roll a more sophisticated enterprise solution. Should be good enough for home disk to disk backups, right?

Mixolyde wrote:

My company is using Synctoy as a stop gap for local backups until we roll a more sophisticated enterprise solution. Should be good enough for home disk to disk backups, right?

One thing to be careful of: apparently Synctoy just mirrors actions from disk A to disk B.

So, if something gets corrupted or deleted on disk B, the "good" copy on disk A isn't going to copy over, because nothing has changed on that file in disk A. It only cares about changes (updates/deletions/etc) that occur on disk A and replicating them on disk B.

At least that's what that blog entry is indicating. No first-hand experience with Synctoy myself.

Still, any backup is much better than no backup, but I would investigate to see if this behavior is true, and if so, I'd probably look for that more sophisticated solution sooner rather than later.

(To flog the whole CrashPlan thing again: you can use CrashPlan for free to make automated encrypted backups between your local machines.)

So ever since I visited the CrashPlan website at work, the overwhelming majority of banner ads and YouTube ads here have been for CrashPlan. How does it know I haven't installed it yet? Cos it better know as soon as I do install it, otherwise that will get obnoxious.

I played around a bit with crashplan, but is there any other app besides their own which can open their archives? I guess that's their business (although I'd say they're selling the online backup service more than locking you into their client), but I'd prefer something a bit more plain I guess.

Scratched wrote:

I played around a bit with crashplan, but is there any other app besides their own which can open their archives? I guess that's their business (although I'd say they're selling the online backup service more than locking you into their client), but I'd prefer something a bit more plain I guess.

Jungle Disk publishes their encryption code, so that if Jungle Disk died, you could utilize that code to decrypt archives (which are stored in your S3 account).

Box.com are giving away twenty five gigabytes of storage for free.

Having read more about crashplan that looks to be the way to go. Set up the original link here and then move it to the friends house for the deltas. Of course we're going to need to up our cable speed to cope with the increase in data exchange since we only get 100GB a month unless you are on their two highest speed plans.

For what it's worth, there's a CrashPlan app for iOS where you can access files you've archived. Since I back up our photo library, this means I can actually view all the photos on my phone.

Any variations on the suggestions for this use case?
- My family has multiple machines and multiple photographers who initially load their images to those machines
- My laptop preferably can move my photo library (It's now have the size of my whole hard drive) out.
- I want to back up those photo libraries over the network (I can keep a machine online)
- I'm willing to get a Drobo or equivalent.
Thanks

(edited with laptop item)

Since you looked up all the Synologies, do you have a recommendation on a home one?

For SOHO use, I'd suggest the DS411j for under $400 and the biggest disks you can afford.

http://www.amazon.com/Synology-DiskS...

These 2tb are on the compatibility list and only ~$100 each:
www.amazon.com/Seagate-ST2000DM001-B...

You could start with 2 and have 4tb raw for about $600 and eventually get to 16tb raw if you need that much.

A new entry into the backup scene: http://aws.amazon.com/glacier/

Not insanely expensive to store, but a restore will cost you a bloody mint. 12 cents per gig for the first 40TB.

I've got a 6TB fileserver at home, so backing that up looks like it would cost about $60/mo, and then if I needed to do a full restore, that would be $720ish.

Malor wrote:

Not insanely expensive to store, but a restore will cost you a bloody mint. 12 cents per gig for the first 40TB.

I've got a 6TB fileserver at home, so backing that up looks like it would cost about $60/mo, and then if I needed to do a full restore, that would be $720ish.

With 6tb you've got to be storing movies? I've got 15tb myself which we use to stream our movie collection around the house. As I tell everyone doing this, why do you need to back those up? Your physical disks _are_ the backups. Sure it'll take time to re-rip them if something happens but so would downloading it all from any cloud service.

Well, the only way I'd really want to do a cloud backup is as a large, encrypted, single volume. I can't easily splice out the things I can quickly restore from the things that are irreplaceable, and still have it be encrypted with MY encryption, as opposed to theirs.

I suppose I could buy some more drives to set up another encrypted backup volume, but if I'm going to do that anyway, I might as well just do them in a USB3 enclosure, and then take them offsite.